Integrations · 26 and counting
Connect the tools you already use.
Code platforms, issue trackers, coding agents, AI models and agent tools. Connect each one once per workspace, and every agent can use it, under the same rules.
26 integrations
Code platforms
Where your repositories live. Specs sync from them and pull requests go back to them.
- GH
GitHub
Code platformsRepositories, issues and pull requests. Install as a GitHub App in a few clicks.
Source controlIssuesPull requestsAutomation triggersSetup guideTrigger events (6)Hide events
issues.labeledissue_comment.createdpull_request.openedpull_request_review.submittedpushcheck_suite.completed
- GL
GitLab
Code platformsProjects, issues and merge requests, on gitlab.com or your self-managed instance.
Source controlIssuesMerge requestsAutomation triggersSetup guideTrigger events (5)Hide events
issue.openissue.updatemerge_request.openmerge_request.updatemerge_request.merge
- AZ
Azure DevOps
Code platformsAzure Repos and Azure Boards: repositories, work items and pull requests.
Source controlWork itemsPull requestsAutomation triggersSetup guideTrigger events (5)Hide events
workitem.createdworkitem.updatedpull_request.openedpull_request.updatedpush
- GIT
Local git
Code platformsRepositories on disk, no code platform needed. Runs commit to their own branches for you to review and merge.
Source controlOffline
Issue tracking
Turn issues into specs and start work when they change.
- JI
Jira
Issue trackingImport labeled issues as specs and start work when issues change. Jira Cloud, Server and Data Center.
IssuesIssue to specAutomation triggersSetup guideTrigger events (2)Hide events
issue.createdissue.updated
- GI
GitHub Issues
Part of the GitHub integrationLabel an issue to hand it to an agent, or turn it into a spec in one click.
IssuesIssue to specAutomation triggers - GLI
GitLab Issues
Part of the GitLab integrationStart work from new and updated issues, and draft specs from them.
IssuesIssue to specAutomation triggers - AB
Azure Boards
Part of the Azure DevOps integrationStart work from new and updated work items, and draft specs from them.
Work itemsIssue to specAutomation triggers
Coding agents
The agent that does the work. Pick one per agent, switch any time.
- OC
OpenCode
Coding agentsAn agentic coder that reads, edits and runs commands in the checkout, in its own container per run, on any model.
AgenticAny model - CA
Claude Agent SDK
Agent SDK workerRun Anthropic's agent SDK and CLI as a Structrl agent, inside the same harness and sandbox.
AgenticAgent SDK - CX
Codex CLI
Agent SDK workerRun OpenAI's coding CLI as a Structrl agent, inside the same harness and sandbox.
AgenticAgent CLI - SDK
Your own agent
Agent SDK workerAny agent SDK or CLI that edits a repository. Adding one is configuration, not code.
Agent SDKAgent CLI - MW
Built-in model worker
Coding agentsOne model request that proposes file edits. Light and fast, for small, well-specified changes and small local models.
Built inAny model
AI models
Bring your own keys, or run models on your own hardware.
- OR
OpenRouter
AI modelsHundreds of models from every major lab behind one API key, including Claude, GPT, Gemini, Llama, Mistral, DeepSeek and Qwen.
Bring your own keyCloud models - API
OpenAI-compatible API
AI modelsAny endpoint that speaks the OpenAI API, in the cloud or in your own network.
Bring your own keySelf-hosted models - AO
Azure OpenAI
OpenAI-compatible APIModels deployed in your own Azure subscription and region.
Bring your own keyCloud models - OA
OpenAI
OpenAI-compatible APIGPT models straight from OpenAI with your own API key.
Bring your own keyCloud models - OL
Ollama
OpenAI-compatible APIOpen models on your own machine or GPU server. Works fully offline.
Local modelsOffline - LM
LM Studio
OpenAI-compatible APIRun open models on your desktop and point Structrl at its local server.
Local modelsOffline - VL
vLLM
OpenAI-compatible APIHigh-throughput serving for open models on your own GPU cluster.
Self-hosted modelsOffline
Agent tools
Give agents the tools and context your team already uses.
- MCP
MCP servers
Agent toolsConnect any Model Context Protocol server over Streamable HTTP or SSE. Its tools become actions for agents and workflows.
Agent actionsWorkflow actions - CLI
structrl CLI
Agent toolsStart and follow runs from CI, scripts or your terminal, signed in with an access token.
AutomationCI/CD - KN
Knowledge
Agent toolsArchitecture decisions, conventions and docs that agents draw context from on every run.
Agent context
Spec formats
Keep writing requirements the way you already do.
- ST
Structrl specs
Spec formatsMarkdown specs with acceptance criteria in .structrl/specs, versioned with your code.
Native - SK
GitHub Spec Kit
Spec formatsSpec Kit specs (specs/NNN-name/spec.md), read as they are.
Import - KI
Kiro specs
Spec formatsKiro requirements (.kiro/specs/name/requirements.md), read as they are.
Import
Your AI engineering team starts with one issue.
Download Structrl free, connect a repository and your own model, and run your first agent today.